BACKGROUND OF THE INVENTION
[0001] This invention relates to microwave power sensors and specifically to a power sensor
which utilizes diodes to sense the power of microwaves that have frequencies near
and above the resonant frequency of the diodes.
[0002] Microwave power sensors have a number of applications in fields ranging from communications
to national defense. Power sensors using diodes as the sensing means have been used
to measure microwave power for a number of years, but the power sensors of the past
could only be used to measure the power of waves having frequencies much below the
resonant frequency of the diodes.
[0003] To illustrate the concept of a resonant frequency, a commonly accepted model of a
diode formed on a semiconductor substrate is shown in Fig. 1. The diode itself is
represented by a junction capacitance C
j and a non-linear resistance R
o. The spreading resistance R
s is the resistance encountered by the current as it flows through the bulk of the
semiconductor substrate. The lead inductance L
i is the inherent inductance of the electric connection to the diode, and the parasitic
capacitance C
p is the capacitance associated with the semiconductor substrate itself found between
the two terminals A and B. Due to the reactive elements C
j, L
i, and C
p, there is a certain frequency at which the diode will resonate, and this resonant
frequency is determined by the values of these elements. Even with today's fabrication
technology, it is only possible to produce diodes having resonant frequencies of as
high as about 60 GHz. The resonant frequency of a diode is important because it determines
the frequency range within which the diode may usually be used as a power sensor.
[0004] A typical power sensing circuit using a diode is depicted in Fig. 2 comprising a
matching load resistor R
m, a power sensing diode D, and a very large output capacitor C
o, the capacitor C
o having a negligible RF impedance. A qualitative plot of the DC output of this sensing
circuit against frequency is provided in Fig. 3 for input waves having the same power
but different frequencies. For input waves having a frequency much below the resonant
frequency of the diode, the impedance of diode D is much greater than R
m so that the input wave 20 encounters a parallel combination of the matching load
resistor R
m and a virtual open circuit due to the large impedance of the diode D. As a result,
most of the power of the wave 20 is absorbed by the matching resistor R
m. The diode D detects the voltage of the input wave 20 at node N₁ and rectifies it
to produce a DC signal across output capacitor C
o which has an amplitude proportional to the square of the voltage across terminals
T₁ and T₂. Since power is proportional to the square of the voltage, the amplitude
of the DC output 22 is thus proportional to the power of the input wave 20. For this
reason, the diode is said to be operating under a square law regimen. The DC output
curve of the circuit of Fig. 2 for input waves having frequencies much below the resonant
frequency of the diode D is depicted by curve portion 24 of Fig. 3. For frequencies
below frequency F
L, the DC output for input waves having the same power is constant, meaning that there
is no frequency dependence. The flatness of curve 24 is quite desirable since the
sensor is a power sensor, not a frequency sensor. Waves having the same power should
cause the sensor to output the same signal regardless of the frequency of the wave.
[0005] For frequencies above F
L, however, the DC output is no longer frequency independent as shown by curve 26.
This is due to the fact that the diode D of Fig. 2 in reality behaves as the circuit
of Fig. 1, which should now replace D in Fig. 2. The DC voltage that appears across
the output capacitor C
o really is proportional to the square of the fraction of the input RF voltage wave
that appears across the parallel combination of C
j and R
o. For frequencies above F
L the diode D behaves very nearly as a series resonant circuit, the resonance being
caused mainly by the inductance L
i and the junction capacitance C
j. As frequency increases and approaches the resonant frequency of the diode, the voltage
across the junction capacitance C
j increases due to the resonance effect, and the DC voltage across the output capacitor
C
o, which is proportional to the square of the voltage across C
j, also increases. Another effect is that, as the frequency of the input signal increases
and approaches the resonant frequency of the diode, the diode impedance decreases,
causing the parallel combination of its impedance with that of the load R
m to present a mismatched load to the input wave. This process continues as the frequency
of the input continues to increase, until the resonant frequency F
R is reached. At this frequency F
R, maximum DC output signal across the output capacitor C
o occurs. The diode impedance reaches a minimum at frequency F
R. As the frequency of the input signal continues to increase, the RF voltage across
the diode junction capacitor C
j starts to drop, and consequently, so does the DC voltage across the output capacitor
C
o. The DC output of the sensor quickly falls off with increased frequency, as shown
by curve portion 28 of Fig. 3. The diode overall impedance begins to rise for frequencies
above F
R. In regions 26 and 28, the sensing circuit of Fig. 2 is outputting different DC signal
levels at different frequencies even though the power of the input wave is constant.
This frequency dependence is quite undesirable. A flat response like that of curve
24 would be more preferable, if not essential. Due to this frequency dependence, diodes
have seldom been used as power sensors at frequencies near or above their resonant
frequencies. As previously mentioned, it is only possible, with current technology,
to produce diodes having resonant frequencies of up to about 60 GHz. However, there
are many applications which require a power sensor capable of operating at frequencies
way above 60 GHz (e.g., between 75 and 110 GHz). Therefore, a need exists for an apparatus
which would allow a diode to be used near and above its resonant frequency to measure
accurately RF power.
[0006] Thus, it is an object of the invention to provide a power sensor which employs a
diode to measure accurately the power of electromagnetic waves having frequencies
at, near, or above the diode's resonant frequency.
[0007] Another object of the invention is to provide a broad-band power sensor which has
a relatively flat frequency response up to between 75 and 110 GHz and above.
[0008] Yet another object of the invention is to provide a power sensor which has a load
impedance which matches the characteristic impedance of the input waveguide.
SUMMARY OF THE INVENTION
[0009] In accordance with the objects of the invention, a broad-band electromagnetic wave
power sensor is provided which is capable of operating at frequencies near and above
the resonant frequency of the sensing diode. The power sensor of the invention comprises
a sensing diode, an input wave conditioning means, and a waveguide. The sensing diode
of the invention is a diode of regular construction having frequency characteristics
such that, for input waves having frequencies near or above its resonant frequency,
the ratio of the DC output of the diode to the power of the input wave (DC output/power
input ratio) is frequency dependent.
[0010] The conditioning means, adapted to receive an input electromagnetic wave and output
a conditioned wave having a fraction of the power of the input wave, is designed in
such a way that it offsets the frequency dependence of the diode. This may be achieved
by designing the conditioning means to transmit a greater fraction of the input wave
to the diode at frequencies where the DC output/power input ratio of the diode is
below a reference value. Likewise, the conditioning means transmits a lesser fraction
of the input wave to the diode at frequencies where the DC output/power input ratio
is above a reference value. In this manner, the DC output of the diode may be kept
relatively constant for input waves having equal power but different frequencies.
The conditioning means is further adapted to have a load impedance which substantially
matches the characteristic impedance of the input electromagnetic waveguide so that
the conditioning means minimizes the reflection of the input wave. This allows the
power of the wave to be measured more accurately. Furthermore, the conditioning means
is adapted to attenuate the input signal to a sufficient degree to ensure that the
sensing diode functions within the square law region.
[0011] The waveguide of the invention receives the conditioned wave outputted by the conditioning
means and conveys it to the sensing diode. The sensing diode responds to the conditioned
wave by outputting a DC signal which is proportional to the power of the conditioned
wave. The sensing diode is preferably attached to the waveguide at a point where a
voltage maximum of the conditioned wave always occurs so as to improve the sensitivity
of the sensing diode across a broad frequency band.
[0012] By combining the conditioning means, waveguide, and diode into a single system, a
power sensor may be produced which is capable of providing very nearly constant DC
output signals for input waves having equal power but different frequencies. Thus,
a broad-band power sensor is provided which may be used to measure the power of electromagnetic
waves having frequencies near and above the resonant frequency of the sensing diode.
BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Fig. 1 is a circuit diagram of a commonly accepted model of a sensing diode constructed
on a semiconductor substrate.
[0014] Fig. 2 is a diagram of a prior art power sensing circuit comprising a matched load
and a sensing diode.
[0015] Fig. 3 is a plot of the DC output of the sensing circuit of Fig. 2 with respect to
the frequency of input waves having equal powers.
[0016] Fig. 4 is a block diagram representation of the basic components of the power sensor
of the invention.
[0017] Fig. 5 is an enhaced view of curves 26 and 28 shown in Fig. 3 to more clearly illustrate
the purpose of the conditioning means of the invention.
[0018] Fig. 6 is a depiction of the desired frequency curve of the conditioning means of
the invention above the diode's resonant frequency.
[0019] Fig. 7 is a plot of the DC output of a typical sensing diode versus RF power input.
[0020] Fig. 8 is an illustration of a prior art waveguide which attempted to place the sensing
diode at a position along the waveguide at which a voltage maximum of the input wave
occurred.
[0021] Fig. 9 is a plot of the DC output/power input ratio of the sensing diode used in
the specific implementation of the invention.
[0022] Fig. 10 is a cross sectional view of a prior art directional coupler used as a signal
separator and sampler.
[0023] Fig. 11 is a top cross sectional view of the prior art directional coupler shown
in Fig. 10.
[0024] Fig. 12 is a qualitative representation of the frequency response of a typical prior
art directional coupler.
[0025] Fig. 13 is a qualitative representation of the desired frequency response of the
coupler of the specific implementation.
[0026] Fig. 14 is a quantitative plot of the actual frequency response of the directional
coupler of the specific implementation to illustrate the invention.
[0027] Fig. 15 is a cross sectional view of the directional coupler of the specific implementation
emphasizing the load of the coupler having an impedance which matches the characteristic
impedance of the input wave.
[0028] Fig. 16 is a cross sectional of the tapered waveguide of the specific implementation
to illustrate the invention.
[0029] Fig. 17a is a frontal view of the input port of the waveguide of Fig. 16 showing
the dimensions of the input port.
[0030] Fig. 17b is a frontal view of the output port of the waveguide of Fig. 16 showing
the dimensions of the output port.
[0031] Fig. 18 is an illustration showing the sensing diode attached to the output port
of the waveguide of the specific implementation to illustrate the invention.
[0032] Fig. 19 is an illustration of the overall power sensor of the specific implementation
of the invention to illustrate the invention.
[0033] Fig. 20 is a normalized quantitative plot of the frequency response of the power
sensor of Fig. 19 to illustrate the relatively flat frequency response of the power
sensor of the specific implementation of the invention.
[0034] Fig. 21 is an illustration showing two sensing dioes whose terminals of opposite
polarity are connected to the output port of the waveguide of an alternative implementation
to illustrate the invention.
[0035] Fig. 22 is a block diagram of a power sensor, where the sensor may be calibrated
for accurate measurement.
[0036] Fig. 23 is a schematic circuit diagram illustrating in more detail the dual diode
sensor and reference attenuator of Fig. 22.
DETAILED DESCRIPTION OF THE INVENTION
[0037] A block diagram illustrating the basic components of the power sensor of the present
invention is shown in Fig. 4, wherein the sensor 29 comprises a conditioning means
30, a waveguide 32, and a sensing diode 34. The sensor 29 detects the power of an
input electromagnetic wave 36 by first having the conditioning means 30 convert the
input wave 36 into a conditioned wave 38 having a fraction of the power of the input
wave 36. The fraction of the input wave 36 which is transmitted as the conditioned
wave 38 is a function of the frequency of the input wave 36. The conditioned wave
38 emerges from the conditioning means 30 and propagates through the waveguide 32
until it reaches the sensing diode 34. In response to the conditioned wave 38, the
diode 34 outputs a DC signal 40 having a magnitude which is proportional to the power
of the input wave 36. It may seem that the output 40 of the diode is not proportional
to the power of the input wave 36 at all, but is instead, an indication of the power
of the conditioned wave 38. Although the DC output 40 of the diode 34 is a frequency
dependent, non flat function of the power of the conditioned wave 38, the conditioning
means offsets the frequency response of the diode 34 in such a way that the power
of the input wave 36 may be measured accurately. The manner in which this may be accomplished
will be explained in a subsequent section. The specifications and requirements of
each of these components, as well as the function of the entire sensor 29, will now
be discussed in greater detail.
Sensing Diode
[0038] The sensing diode 34 of the invention may be any of a number of typical sensing diodes
currently used as power sensors. A qualitative plot of the DC output of these typical
diodes versus the frequency of the input waves is shown in Fig. 3 for input waves
having the same RF power but different frequencies. Fig. 5 provides an enhanced view
of the curve portions 26 and 28 of Fig. 3. According to Fig. 5, input waves having
frequencies near and above the resonant frequency F
R of the diode cause the diode to output DC signals which differ depending upon the
frequency of the input wave. In other words, the ratio of the DC output of the diode
to the power of the input wave (hereinafter referred to as the DC output/power input
ratio) is frequency dependent. For example, the DC output/power input ratio R₁ for
the frequency F₁ is larger than the ratio R₂ for the frequency F₂. This frequency
dependence is quite undesirable because the diode is a power sensor, not a frequency
sensor. Input waves having the same RF power should cause the diode to output the
same DC signal regardless of the frequency of the input wave. It is precisely for
this reason that diodes have not been used effectively in the past as power sensors
for frequencies near and above the resonant frequency of the diode.
Conditioning Means
[0039] The conditioning means 30 of the invention serves a number of different purposes
but none more important than that of offsetting the frequency dependence of the sensing
diode 34. This is accomplished by designing the conditioning means 30 to vary the
fraction (hereinafter referred to as the transmission factor) of the power of the
input wave 36 transmitted to the sensing diode depending upon the DC output/power
input ratio of the diode for a particular frequency. To elaborate, for frequencies
where the DC output/power input ratio is below a reference value, the conditioning
means sends a relatively large fraction of the input wave power to the sensing diode.
For frequencies where the DC output/power input ratio is above a reference value,
the conditioning means transmits a lesser fraction of the input wave power to the
sensing diode. By varying the transmission factor of the conditioning means, the DC
output level of the diode for input waves having equal powers but different frequencies
may be kept nearly constant over a wide frequency range.
[0040] To illustrate this more clearly, reference is made to Figs. 4 and 5. Suppose that
an RF wave having a frequency F₁ is conveyed to the power sensor 29. According to
Fig. 5, the DC output/power input ratio R₁ for this frequency is relatively large.
Therefore, a relatively small fraction of the input wave 36 should be transmitted
to the diode 34 so as to cause the diode 34 to output a DC signal having a magnitude
approximately equal to DC
ref. Suppose instead that another RF wave having the same power as the first but at a
frequency of F₂ is conveyed to the power sensor 29. For the frequency F₂, the DC output/power
input ratio R₂ is relatively small so that a relatively large fraction of the input
wave 36 should be transmitted to the diode 34 to cause the diode to output a DC signal
whose magnitude is once again approximately equal to DC
ref. By properly varying the transmission factor of the conditioning means for each frequency,
the diode may be caused to output a fairly constant DC voltage (DC
ref) over a wide range of frequencies (from F₁ to F₂). This is illustrated by the line
42 of Fig. 5. In order to accomplish this, the transmission factor of the conditioning
means should increase with increasing frequency. A qualitatively plot of the transmission
factor of the conditioning means versus frequency is shown in Fig. 6. The slope of
the transmission factor curve should be designed to specifically offset the downward
sloping response (curve 28 of Fig. 2) of the sensing diode to cause the diode to produce
a relatively constant DC output response for input waves having equal power but different
frequencies.
[0041] In addition to compensating for the frequency dependence of the sensing diode, it
is very important that the conditioning means 30 be further adapted to provide a matching
load impedance for the input wave. By having a load impedance that matches the characteristic
impedance of the input wave, the conditioning means minimizes the reflection of the
input wave, thereby, allowing the total power of the input wave to be measured. This,
in turn, allows the sensor of the invention to very accurately measure the power of
the input wave.
[0042] Yet another function of the conditioning means is to attenuate the input wave to
ensure that the sensing diode functions in the square law region when it senses the
conditioned wave. With reference to Fig. 7 which shows a plot of the voltage output
of the diode versus the RF power sensed by the diode, the diode has a linear region
44 of operation and a non-linear region 46 of operation in power terms. In terms of
voltage, the linear region 44 is referred to as the square law region because the
DC output of the diode in this region is directly proportional to the RF power sensed
by the diode. Beyond an RF power of approximately -20dBm, however, the diode begins
to saturate and its output voltage no longer linearly follows the input RF power.
To ensure that the diode functions within the linear region, the conditioning means
30 of the invention preferably attenuates the input wave to such a degree that the
conditioned wave sensed by the diode has an RF power of -20dbm or less.
Waveguide
[0043] The waveguide 32 of the invention is preferably a tapered, open-ended, rectangular
transmission line with a larger impedance at the input port than at the output port.
The impedance of the transmission line is reduced from the input port to the output
port by tapering the height of the transmission line but leaving the width unchanged.
The impedance,of the waveguide at the open output port should be sufficiently small
compared to the impedance of free space such that the transition from the waveguide
to free space is sufficiently abrupt to create a virtual open circuit. This open circuit
causes the conditioned wave to have a voltage maximum at the output end of the waveguide,
and this voltage maximum occurs there regardless of the frequency of the conditioned
wave. The physical dimensions of the open output port of the waveguide, specifically
the height, also make it possible to install the diode while minimizing the parasitic
effects of its connections.
[0044] This is in sharp contrast to the waveguides used in prior art power sensors. A cross-sectional
view of one such waveguide is shown in Fig. 8 wherein the rectangular waveguide 50
has a sensing diode 52 and a tuning element 54 attached thereto. The sensor of Fig.
8 operates in a useful manner only at one frequency, or, at most, over a very narrow
range of frequencies centered around the frequency for which the tuning element is
adjusted. As the frequency of the input wave changes, in order to measure the power
associated with it the tuning element must be re-adjusted and furthermore, the sensor
must be recalibrated. This process can be quite tedious and impractical, especially
if the frequency of the input wave 56 is frequently changed.
[0045] In sharp contrast, the waveguide of the present invention creates a virtual open
circuit at the output port of the waveguide, which in turn causes a voltage maximum
of the input wave 56 to always appear there regardless of the frequency of the input
wave. Therefore, overall sensor sensitivity is kept at a maximum. Additionally, the
match that the diode and the output waveguide port present to the conditioned wave
is irrelevant to the match of the sensor, which is solely determined by the conditioning
means 30 of Fig. 4. As a result, the sensor 29 of the invention may be used to effectively
measure the power of input waves over the full nominal frequency band of the standard
waveguide to which it is connected.
[0046] Thus far, only the general specifications and requirements of the elements of the
invention have been disclosed. To more clearly illustrate the invention, a specific
implementation of the invention will now be discussed.
Specific Implementation
[0047] It is the goal of this specific example to provide an RF power sensor which is capable
of accurately sensing the RF power of waves having frequencies way above 60 GHz, e.g.,
between 75 and 110 GHz.
Sensing Diode
[0048] The sensing diode used in this example is a typical low barrier solid state diode
which has a resonant frequency of approximately 60 GHz. A plot of the DC output/power
input ratio of this diode versus frequency is shown in Fig. 9. From Fig. 9, it is
clear that the DC output/power input ratio of the diode falls off significantly with
increasing frequency. The input mismatch of such a sensor is also severe. This is
to be expected since the frequency range of interest is significantly above the resonant
frequency of the diode.
Directional Coupler
[0049] As previously discussed, the conditioning means of the invention serves three major
purposes: (1) offsetting the frequency dependence of the sensing diode; (2) providing
a matching impedance for the input wave; and (3) attenuating the input wave sufficiently
to ensure that the sensing diode operates within the square law region. These functions
are preferably implemented through the use of a properly designed directional coupler
but not restricted to the use of such couplers (other circuits will also work). Before
describing the directional coupler of the invention, however, a general discussion
on directional couplers is first necessary.
[0050] A cross sectional view of a typical directional coupler is shown in Fig. 10 wherein
the coupler 60 comprises a main waveguide 62 and an auxiliary waveguide 64, the two
waveguides 62, 64 being separated from each other by a conductive wall 66 having a
thickness T. An input wave 70 may enter the coupler 60 through the main input port
72 of the main waveguide 62 and propagate through the main waveguide 62 to the main
output port 76. The holes 68 in the separating wall 66 allow a portion of the input
wave 70 to enter the auxiliary waveguide 64 and propagate through to the auxiliary
output port 74. Because coupler 60 separates the input signal 70 into more than one
output signal, directional couplers like coupler 60 are usually used as signal separators
and samplers.
[0051] With reference to Fig. 11, which provides a top cross sectional view of the coupler
of Fig. 10, the separating wall 66 of the coupler 60 contains a plurality of holes
68 which allow portions of the input wave 70 to enter and propagate through the auxiliary
waveguide 64 (Fig. 10). The fraction of the wave 70 that enters the auxiliary waveguide
64 is defined as the coupling coefficient of the coupler 60 and usually varies with
the frequency of the input wave 70. Directional couplers using multiple hole coupling
are known. See "A Precision Directional Coupler Using Multi-Hole Coupling," by Barnett
et al., Hewlett-Packard Journal, Vol. 3, No. 78, March-April 1952.
[0052] With reference to Figs. 10 and 11, the coupling coefficient of the coupler 60 is
a function of the following parameters: (1) the number of holes 68 and the diameter
D of each of the holes 68 in the separating wall 66; (2) the thickness T of the separating
wall 66; (3) the distance X₁ between the side of the coupler 60 and the center of
the holes 68; (4) the width W of the coupler 60; (5) the height H of the main and
auxiliary waveguides 62, 64; and (6) the wavelength λ
g of the input wave 70 within the waveguides 62 and 64. The magnitude of the coupling
coefficient for each of the holes 68 in the separating wall 66 may be expressed by
the following equation:

where




and

For most typical couplers, the constant C ranges from .87 to .98 and the constant
K ranges from 2.8 to 3.9. The coupling coefficient of the entire coupler 60 may be
obtained by summing up the coupling coefficients of each of the individual holes 68.
[0053] The attenuation and frequency characteristics of a typical directional coupler is
that shown in Fig. 12 wherein the attenuation versus frequency curve takes on a bell-like
shape. The attenuation factor of the coupler has a maximum around the middle of the
operative frequency range of the coupler with a swing between the maximum and the
minimum attenuation being within a 1dB window.
[0054] Although there are six parameters which affect the coupling coefficient of the coupler
60, there are three which are of particular significance. These parameters are the
number of holes 68 and the diameters of the holes 68 (Fig. 10), the thickness T of
the separating wall 66 (Fig. 9), and the distance X₁ (Fig. 10). The attenuation factor
of the coupler is mostly affected by the diameters of the holes 68 and the thickness
of the separating wall 66, while the frequency shaping of the coupler is mainly determined
by the distance X₁ between the center of the holes 66 and the side of the coupler
60. By making the other parameters constant and by only manipulating these three parameters,
a coupler may be designed which meets the specifications for the conditioning means
of the invention.
[0055] With reference to Fig. 12, the frequency response of a typical directional coupler
clearly does not meet the specifications of the conditioning means of the invention.
Instead of having a bell-like shape, the directional coupler of the invention should
have an attenuation versus frequency curve which resembles that depicted in Fig. 13.
The bell-like shape of the coupler's frequency response may be altered, however, by
adjusting the distance X₁ (Fig. 10). As shown in Fig. 12, a proper adjustment in X₁
will pull the attenuation factor at frequency F₁ up and push the attenuation factor
at frequency F₂ down so that the coupler's frequency curve takes on a shape which
does resemble that shown in Fig. 13.
[0056] At first glance, it may seem that the desired attenuation/frequency curve should
be one that slopes upward, not downward as depicted in Fig. 13. It is true that the
curves in Figs. 6 and 13 do not seem to agree at first but they actually are conveying
the same message. With reference to Fig. 13, as the frequency increases, the attenuation
factor of the coupler decreases. That is, as the frequency of the input wave increases,
a greater fraction of the input wave is being transmitted by the coupler. Therefore,
the transmission factor of the coupler increases with frequency just as is depicted
in Fig. 6. Thus, in essence, the curves in Figs. 6 and 13 are equivalent.
[0057] With all of these factors in mind, the directional coupler of this specific implementation
was designed as follows:
(1) thickness T of the separating wall = .254 mm (millimeters);
(2) five pairs of holes were placed in the separating wall with the diameters of these
holes being .311, .294, .262, .218, and .166 mm;
(3) the distance X₁ = .665 mm;
(4) the height H of the main and auxiliary waveguides = 1.270 mm; and
(5) the width W of the coupler = 2.540 mm.
A plot of the attenuation factor versus frequency of a directional coupler having
these values is shown in Fig. 14. This directional coupler does indeed have the desired
frequency characteristics.
[0058] As thus far designed, the directional coupler satisfies two of the three criteria
of the conditioning means, namely the attenuation and the frequency dependence compensation
criteria. The third criterion still to be met is the impedance matching criterion.
This is easily accomplished, however, as shown by Fig. 15. Instead of leaving the
output port 76 of the main waveguide 62 open as shown in Fig. 10, the main waveguide
of the directional coupler 79 of this implementation is terminated with a load 80
having an impedance which is substantially equal to the characteristic impedance of
the input wave. As far as the input wave 70 is concerned, when it propagates into
the main waveguide 81, it encounters a matched load 80. As a result, very little of
the input wave 70 is reflected and this, in turn, allows the power of the wave to
be sensed more accurately. With the main waveguide terminated with a matching load,
the directional coupler only has one output port, the auxiliary output port 83. Thus,
it is this port which transmits the conditioned wave to the waveguide.
Waveguide
[0059] It was previously mentioned that the waveguide should be tapered to reduce the height
of the output port so that the impedance at the input port is larger than the impedance
at the output port. Such a waveguide is shown in Fig. 16 wherein the waveguide 86
has an input port 87 for receiving a conditioned wave 88 and an output port 89 for
transmitting an output wave. With reference to Figs. 17a and 17b, the input port 87
has a width W₁ equal to about .100 inches and a height of about .05 inches. In contrast,
the output port 89 has the same width W₁ as the input port 87 but a height H₂ of only
about .005 inches. The reduced height of the output port 89 significantly reduces
its impedance. Thus, when the wave 88 propagates to the output end 89 of the waveguide
86, it encounters a sudden and tremendous impedance difference between the small impedance
of the output port 89 and the relatively large impedance of free space. As a result,
a virtual open circuit exists at the output port/free space boundary. Because of the
open circuit, a voltage maximum of the wave 88 will always appear at the output port
89 of the waveguide regardless of the frequency of the wave. Because this is true
and because it is preferable for the sensing diode to sense the wave 88 at a voltage
maximum of the wave 88, the sensing diode of this implementation should be connected
across the output port 89 of the waveguide 86 as shown in Fig. 18. This configuration
also reduces the parasitic effects of the diode connections to a minimum.
[0060] With reference to Fig. 18, the sensing diode 90 is mounted across the open output
port 89 of the waveguide 86 with each of the diode terminals attached to a corresponding
MIS capacitor 92a and 92b (approximately 2pF). A wire 93 connects capacitor 92a to
the waveguide 86 to ground one of the terminals of the diode 90. Another wire 94 connects
the output terminal of the diode 90 to a filtering capacitor 95 (approximately 1000pF)
to allow the DC output of the diode to be taken from the filtering capacitor 95. Being
thus attached, the sensing diode 90 will always sense the power of the wave 88 at
a voltage maximum regardless of the frequency of the wave 88. Therefore, the diode
90 may be used to sense the RF power of electromagnetic waves over a broad frequency
band.
Complete Power Sensor
[0061] With all of the elements properly designed, the overall power sensor may now be assembled
as shown in Fig. 19 wherein the sensor 100 comprises the directional coupler 79, and
the tapered waveguide 86 with the sensing diode 90 attached to its output port. The
DC output versus frequency response of this sensor is shown in Fig. 20. The plot of
Fig. 20 is normalized by dividing all of the DC output values by the maximum DC output
value. Thus, the values on the vertical axis are percentages of the maximum DC output
value. From Fig. 20, it can be seen that the frequency response of the overall power
sensor is indeed relatively flat over the entire frequency band from 75 Ghz to 110
GHz.
[0062] It should be noted at this point that the plot of Fig. 20 is the calibration curve
for a sensor of this invention. It reduces to a minimum the errors due to sensor lack
of absolute flatness as a function of frequency, as is customary in the art. As long
as the frequency of the input wave can be ascertained, which is normally the case,
the apparatus of the invention may be used effectively to sense the power of electromagnetic
waves having frequencies near and above the resonant frequency of the sensing diode.
[0063] Fig. 21 is an illustration showing two sensing diodes whose terminals of opposite
polarity are connected to two output ports of the waveguide of an alternative embodiment
to illustrate the invention. To simplify the description, identical components are
labeled with the same numerals. As known to those skilled in the art, thermal effects
on the sensing diodes such as diode 90 in Fig. 19 may cause the DC output voltage
of the diode to be in error. To compensate for such thermally induced errors, an additional
sensing diode 90′ is employed as shown in Fig. 21, where the second diode is also
connected across the open end of the output port 89 of the waveguide 86. In the manner
similar to connecting the arrangement for sensing diode 90, a wire 93′ connects capacitor
92a′ to the waveguide 86 to connect to ground one of the terminals of the diode 90′.
Another wire 94′ connects the output terminal of the diode 90′ to a filtering capacitor
95′ (again approximately 1,000 pF) to allow the DC output of the diode to be taken
from the filtering capacitor 95′. As shown in Fig. 21, the positive terminal of diode
90 is connected by wire 94 to capacitor 95 whereas for diode 90′, the negative terminal
is connected by wire 94′ to filtering capacitor 95′. The two outputs at capacitors
95, 95′ are then both fed to a balanced chopper amplifier as shown in Fig. 22. In
this manner, error voltages caused by thermal effects on the sensing diodes are cancelled.
[0064] Fig. 22 is a block diagram of a power sensor to illustrate another embodiment of
the invention. As shown in Fig. 22, power sensor 150 includes a 50 MHz reference attenuator
152 which has a reference calibration signal input 154. Because of variations in the
gain and sensitivity of the different components in a power sensor 150, in order for
power meter 156 to provide an accurate measurement, the power sensor 150 is preferably
first calibrated before the power measurement is taken. For this purpose, when no
signal is fed to the RF input 158, a reference calibration signal such as one at 50
MHz at 1 milliwatt is applied to reference input 154. The two outputs of the dual
diode sensors 90, 90′ are fed to the balanced chopper amplifier 162 where the amplified
diode power output is read by power meter 156 for the purpose of calibrating the meter.
Typically, the gain of the meter is adjusted so that the power reading is at a given
level such as 1 milliwatt corresponding to a 50 MHz input reference calibration signal
at input 154. The power sensor 150 is then ready to be used for measuring a radio
frequency signal of an unknown frequency applied to input 158. Hewlett-Packard Company
of Palo Alto, California, the assignee of this application, provides many instruments
which provide a 50 MHz calibrating signal where the instruments have also been measured
in the factory for various offsets and other frequencies by reference to a 50 MHz
reference signal. The above-described power sensor 150 is therefore particularly convenient
when used in conjunction with such instruments. For a description of the performance
and design of power sensors provided by Hewlett-Packard Company, see "Design and Performance
of Millimeter-Wave Thermocouple Sensors," by Lee H. Colby, Hewlett-Packard Journal,
pp. 35-37, April 1988.
[0065] Fig. 23 is a schematic circuit diagram illustrating in more detail the dual diode
sensor 160 and the reference attenuator 152 of Fig. 22. As shown in Fig. 23, diode
90 has an input 90(i) and an output 90(o). Diode 90′ has an input 90′(i) and an output
90′(o). As shown in Fig. 23, the reference attenuator 152 connects reference input
154 to the two inputs of the two diodes 90, 90′, where the outputs of the two diodes
are connected to the balanced chopper amplifier shown in Fig. 22. In this manner,
a reference calibration signal applied to input 154 will be sensed by both diodes
which provide outputs to the balanced chopper amplifier and measured by power meter
156 as described above.
[0066] Although the invention has been described with reference to a specific example, it
should not be construed to be so limited. Many modifications may be made by one skilled
in the art with the benefit of this disclosure without departing from the spirit of
the invention. For example, the conditioning means need not be implemented with a
directional coupler. Implementations using other means are possible. This and other
modifications are well within the scope of the invention. Thus, the invention should
not be limited by the examples used to illustrate it but only by the scope of the
appended claims.